Tech

Google quantum processor beats classical limits in new test of quantum physics

Researchers demonstrated that a fundamental quantum property called contextuality survives in processors with up to 105 qubits, beating classical limits in complex coordination games.

Marissa Giustina Demonstrating Quantum Supremacy
Google / Wikimedia Commons, CC BY 3.0

Google researchers have demonstrated a key feature of quantum physics known as contextuality at scale, outperforming classical computer limits on tests using up to 105 qubits, according to a paper published in the journal Science.

The study, featured on the cover of Science, provides experimental evidence that contextuality persists in large, multi-particle quantum systems. While quantum entanglement has long been viewed as the primary engine of quantum computing, theoretical physicists increasingly consider contextuality a deeper resource that powers quantum algorithms past their classical counterparts, Google reported.

Challenging Classical Physics

In classical physics, physical properties are assumed to exist independently of whether they are measured. Albert Einstein famously questioned this idea by asking whether the moon exists when no one is looking. Quantum contextuality provides a framework showing that measurement outcomes are not fixed in advance; instead, what a measurement reveals can depend on which other compatible measurements are conducted alongside it.

To test the phenomenon at scale, Google Quantum AI used its superconducting processors, including the Willow chip, to play a mathematical puzzle known as the magic square game. In the game, two players who cannot communicate must fill cells in a three-by-three grid under specific mathematical rules. Under classical mechanics, a pre-planned strategy can win at most eight out of nine rounds, or 88.89 percent.

Using shared entangled qubits, Google’s Willow chip won the game 98.3 percent of the time across nine variations of the challenge, with each version played 100,000 times, MLQ.ai reported. Google noted that hardware noise accounted for roughly one loss in every 60 rounds, though flawless execution remains possible in theory.

Scaling Up Quantum Benchmarks

The research team extended their testing to larger quantum setups. In a Bell-Kochen-Specker contextuality test, where non-contextual systems cannot exceed a score of 4 and quantum physics permits a maximum score of 6, Willow scored 5.590. The state-independent test involved 540 sequential measurements repeated 1,000 times, coming within roughly 7 percent of the theoretical quantum limit.

Google also evaluated performance in a multi-player parity game. The best classical strategy achieves a 75 percent win rate with three players and approaches 50 percent as more players are added. Google’s Willow chip maintained a win rate above the classical threshold with up to 71 players while confirming entanglement across 45 qubits, according to MLQ.ai.

Additionally, on a two-dimensional hidden linear function task, shallow quantum circuits across grids of up to 105 qubits delivered valid answers more than half the time, with success rates dropping from nearly 1.0 down to approximately 0.6 at maximum size.

New Diagnostic Tool for Quantum Hardware

Beyond proving theoretical principles, Google reported that contextuality testing serves as a sensitive new diagnostic tool for quantum computer hardware. The tasks identified subtle hardware errors that standard random-circuit benchmarks failed to catch.

Researchers expect the refined error-detection methods to assist in advancing quantum systems toward practical real-world applications, including complex simulations in chemistry and materials science.